Skip to content
Notifications
Clear all

Has anyone used WellSaid for generating audio for visually impaired users on a website?

1 Posts
1 Users
0 Reactions
0 Views
(@ethanp)
Estimable Member
Joined: 2 weeks ago
Posts: 130
Topic starter   [#22591]

Greetings, community members. This is a query that sits at an interesting intersection of accessibility, user experience, and synthetic voice technology, and I believe it warrants a detailed discussion. While we often examine text-to-speech (TTS) platforms like WellSaid Labs through the lens of content creation, marketing, or e-learning, their application in making web content accessible to visually impaired users presents a distinct set of criteria and potential challenges.

The core question is whether a high-quality, proprietary TTS service like WellSaid is a suitable and practical solution for generating real-time or pre-rendered audio versions of website content. Key considerations here would include the technical implementation, cost structure relative to the scale of content, and most critically, how the output aligns with established web accessibility standards (WCAG). For instance, does the natural cadence and intonation of a WellSaid voice agent enhance comprehension over a standard system voice, and does that justify the operational complexity? Furthermore, one must consider control elements: can a user pause, rewind, or adjust the speech rate, which are common expectations in screen reader software?

I am particularly interested in hearing from members who have implemented or piloted such a system. Practical insights would be invaluable. For example:
* Was the audio generated dynamically via API for each page load, or was it pre-rendered for static content?
* How did you handle frequently updated content, and what was the associated latency or cost implication?
* Did you conduct any user testing with visually impaired individuals to gauge the effectiveness and usability compared to traditional screen readers like JAWS, NVDA, or VoiceOver?
* Were there any specific technical hurdles in embedding the audio players in an accessible manner (e.g., proper ARIA labels, keyboard navigation)?

The philosophical and practical balance here is fascinating—between leveraging cutting-edge, human-like TTS for a better experience and ensuring the solution remains robust, affordable, and, above all, functionally accessible. I look forward to a constructive exchange of experiences and technical perspectives on this nuanced topic.

— EthanP


Let's keep it constructive


   
Quote